I have a preliminary analysis of wishcasting the election. Once it’s finished, I’ll post the entire thing here (and in a paper). Quick conclusion: both sides had their fair share of shading their bets towards their picks.
One thing that’s missing (I ran out of my 750 words) is the proviso that the results are “on average”, which is the assumption that the evidence (let the reader understand) is roughly the same for all participants. This isn’t quite true, but it isn’t far off to the first approximation.
Anyway, head on over and take a read. Let me know what you think.
Update Links fixed. Thanks, DAV!
Both links return to here.
“the chance of fibbing, spoofing, and other forms of dishonesty when answering surveys is not trivial; but I decided, like all sociologists do, to ignore it.”
Nice dig…:-)
Given the results, it is clear that the Mayans were a little off: their calendar ran out on November 6, not on December 21. Despite my wishes to the contrary.
It’d be helpful if you could present two-way or three-way tables of Pick by Want for each gender, and Pick by Party affiliation for each gender.
The cross-classification tables would allow me to answer some questions about your readership and the reliability of your convenient samples.
The empirical probability P( wanting Romney to win | picking Romney to win) of 0.8 is used to define so-called “wishcasting bias†(wb) for Romney. I am not sure exactly what wb measures. If wishcasting is a form of wishful thinking, which is a bias toward a belief due to a given wish/want, then in the election scenario it would seem more appropriate to somehow connect wishcasting to the probability P(picking Romney to win | wanting Romney to win).
Intuitively, if the P(a Romney win)= 0.98, a very high probability, then the P(picking Romney to win) should also be high if there is no wishful thinking involved when picking the winner. Anyway. Will need time to think about it!
JH,
Thanks. All that and more coming. I had 750 words at the LI site. And I used up all of them.